Document not found! Please try again

Using Agricultural Ontologies - Springer Link

8 downloads 23664 Views 101KB Size Report
the construction of an ontology, concerning the horticultural domain of. Agriculture, and ... initiatives towards the construction of agricultural domain ontologies based on .... Simperl, E. P. B., Tempich C.: Ontology Engineering: a Reality Check.
Using Agricultural Ontologies M. T. Maliappis

Abstract. Despite the growing number of ontologies online available, their range of application in real world projects is comparatively limited. This paper describes the construction of an ontology, concerning the horticultural domain of Agriculture, and examine its usage in particular application areas. The proposed ontology is used, (a) as a refine and classification tool facilitating searching process in a repository environment, and (b) as a domain model for rule knowledge base construction. A prototype web application has been developed to investigate the usefulness of ontology and examine several software development tools.

1 Introduction Knowledge collection and representation in a usable and efficient way is a cumbersome, time consuming and expensive process. Sharing of the same knowledge in different applications and its reuse without or little modifications in solving separate problems is a critical factor in knowledge dissemination. Among knowledge representation techniques, ontology formalization can be used to model real world in a consistent, formal, manageable and reusable way. Ontologies can be used to facilitate several aspects of knowledge management such as knowledge representation, knowledge sharing and reuse, knowledge classification and knowledge search and retrieval. There are several efforts towards construction and usage of ontologies in several fields. In Agriculture domain, there is a number of initiatives towards the development of specific ontologies[1,5]. Among the most interesting and promising endeavors is the work carried on in FAO [2]. FAO started several initiatives towards the construction of agricultural domain ontologies based on FAO’s Multilingual Agricultural Thesaurus (AGROVOC) [3,9] and through the Agricultural Ontology Service (AOS) project [4].

Michael T. Maliappis Informatics Laboratory, Agricultural University of Athens, 75 Iera Odos, 11855, Athens, Greece, e-mail: [email protected]

493

494

M. T. Maliappis

Despite the growing number of ontologies online available, their range of application in real world projects is comparatively limited [8]. A major objective of this paper is to construct an ontology targeting horticultural domain of Agriculture and examine its usage in particular application areas. The proposed ontology describes production of horticultural crops in low technology greenhouses and includes several financially important vegetable crops in the area of Mediterranean basin, such as tomato, pepper and aubergine. Among the objectives of the proposed ontology is the systematic organization and representation of knowledge and terminology, concerning all stages of horticultural production and marketing. The developed ontology has been used in two specific applications. In the first, application ontology has been incorporated as controlled vocabulary into DSpace repository. The provided by DSpace indexing mechanism has been modified to use synonyms from ontology structure and to handle Greek language intricacies such as term stemming. Since DSpace uses Apache Lucene search engine as its underlying search and indexing mechanism most of the work targets to modifications of Lucene functions. The second application is trying to incorporate ontology into a traditional rule based expert system. The initial system [7] was a diagnostic expert system which could be used to identify the principal pests, diseases and nutritional disorders of some common vegetable crops and provide guidance for their control in plastic covered greenhouses. Ontology is used to provide the facts and the attributes of the fact-attribute-value structure of the rules.

2 Ontology Construction Usage of traditional IF-THEN rules in expert system applications is a predominant way in knowledge base construction, since they offer a rich expressive environment accompanied with a freedom in expression and design. During development of several knowledge bases and specific expert systems [6, 7, 10] was recognized that formalization and standardization of the needed knowledge is an absolute necessity towards the construction of efficient and robust expert system applications. Lack of formal and disciplined way of knowledge representation made development and maintenance process a difficult task. Starting from these observations, an effort was started to investigate other possible knowledge representations which, keeping the same expressive power, will offer more discipline in knowledge representation and opportunities for knowledge reuse, allowing easy reengineering of expert system applications developed so far. Ontologies were identified as the knowledge structure filling these requirements. OntoCrop ontology was the result of this endeavor. It covers a portion of the horticultural domain and the cultivation of vegetable crops in low-technology

Using Agricultural Ontologies

495

greenhouses. Since ontology construction is an evolutionary process and goes through several stages until its completion, its development started from the part needed for reengineering the already existed applications. Emphasis has been set to a clear and consistent description of the domain under investigation. Other objectives of this attempt were to describe the specific domain taking into consideration the usage of the ontology and to provide enough information to use ontology in a multilateral and multilingual environment. Identification of unique, with well-defined semantics, relationship types between concepts in the investigated domain were another desired goal [9]. At a second stage, ontology augmented with characteristics to assist its usage in organizing and searching for information. For this purpose, ontology enriched with synonyms, and in some cases with antonyms, of the main terms in English and Greek. Multilanguage dimension is an important issue in countries with native languages different than the dominant languages and more specifically than English. So, the capability for knowledge indexing and searching in native language, in parallel with English, is a crucial factor in application usefulness. OntoCrop has been constructed using Protégé2 ontology editor. OntoCrop contains knowledge concerning cultivation techniques, pest management and crops’ physiology. OntoCrop has, also, been extended to include knowledge about propagation, post-harvest physiology, consumption and marketing.

3 Using Ontology in Rule Reasoning OntoCrop ontology has been used to investigate effectiveness of ontology usage in expert system development using IF-THEN rule reasoning structure This application is a reengineering of a previous developed application [7] which was using traditional IF-THEN rules to construct its knowledge base. The initial system was a diagnostic expert system which could be used to identify the principal pests, diseases and nutritional disorders of six common vegetable crops (aubergine, bean, cucumber, lettuce, pepper and tomato) and provide guidance for their control. In its reengineering form the application incorporates OntoCrop in the construction of its rules providing facts and attributes of the fact-attributevalue structure of the rules. Reengineering of the expert system application started from the construction of OntoCrop describing the set of concepts covered by the rules. During the process were identified duplications in concepts and relationships between them. The rules were rewritten using the refined concepts of ontology. Access to ontology contents had been accomplished using Protégé API. Application’s front end accepts input data through special forms, converts them in XML format and forwards them to 2

http://protege.stanford.edu

496

M. T. Maliappis

the server. The reverse process is followed to present the results with system suggestions to the user.

4 Using Ontology in Indexing and Searching The second application uses ontology to assist indexing and searching of information material relevant to crop cultivation domain. Information material is stored in a DSpace3 repository which uses Apache Lucene4 as search engine. DSpace is able to use a controlled vocabulary for indexing and searching purposes. This vocabulary has a specific XML structure in which OntoCrop is converted using Castor software. Having the proper controlled vocabulary the user is able to select terms from this structure to index the submitted material. DSpace provides a reverse mechanism to search the stored material using terms of the controlled vocabulary. Apart from the simple incorporation of OntoCrop as controlled vocabulary into DSpace a major modification of Lucene Java classes is needed to be able to handle synonyms in indexing and searching. Two Java classes of Lucene have been modified to handle customized indexing and searching processes. The class used for indexing modified to be able to search, find and use the synonyms of the terms selected for indexing and handle peculiarities of Greek language. Synonyms can be in the same or different language, English or Greek. A similar modification made to the Java class used for searching to return material corresponding to the synonyms of searching term. Protégé API used to access OntoCrop to get synonyms of specified terms.

5 Conclusion This paper describes the initial endeavors for the development of an ontology concerning the horticultural domain. Through its usage the proposed ontology passed several rounds of refinement. At its current stage of development it contains useful knowledge targeted to several kinds of users. Students can consult the ontology to identify and learn concepts and relationships between them for the horticultural domain. Growers are able to search for useful knowledge concerning specific cultivation techniques in the field. A major objective of this work is the investigation of the usage of the same ontology in different situations and the identification of the difficulties of this endeavor in order to modify ontology structure or to enhance the used software 3 4

http://www.dspace.org http://lucene.apache.org

Using Agricultural Ontologies

497

tools. Ontology was proved a valuable tool in development of rule-based expert systems, offering concrete structure and discipline which are missing from these systems. A guided searching process with the help of domain representation structures, such as domain ontologies, would be able to direct the user to find more clearly and easily the desired information or knowledge and help in proper application of knowledge in the daily activities of farms. Quality, precision and accuracy of the searching results are heavily depended on the quality of domain ontology and indexing process. In the future, OntoCrop ontology is going to be augmented to include more portions of the horticultural domain. Furthermore, an application is going to be developed offering access, searching and navigation, to OntoCrop ontology using web services technology. Acknowledgments This work is supported by the “PYTHAGORAS-II” research project, which is co-funded by the European Social Fund and Greek national resources (EPEAEK II).

References 1. Beck, H. W., Kim, S., Hagan, D.: A Crop-Pest Ontology for Extension Publications. In: Proceedings of 2005 EFITA/WCCA Joint Congress on IT in Agriculture, Vila Real, Portugal (2005) 2. Beck, H., Pinto, H. S.: Agricultural Ontology Service. UN FAO (2002) 3. Fisseha, F.: Towards better Semantic Standards for Information Management AGROVOC and the Agricultural Ontology Service (AOS). UN FAO, Rome, Italy (2002) 4. Hagedorn, K., Fisseha, F.: Agricultural Ontology Service (AOS). A tool for Facilitating Access to Knowledge. UN FAO, Rome, Italy (2001) 5. Koenderink, N. J. J. P., Top, J. L., van Vliet, L. J.: Expert-Based Ontology Construction: A Case-Study in Horticulture. In: 16th International Workshop on Database and Expert Systems Applications (DEXA'05) (2005) 6. Mahaman, B. D., Passam, H. C., Sideridis, A. B., Yialouris, C. P.: DIARES-IPM: a diagnostic advisory rule-based expert system for integrated pest management in Solanaceous crop systems. Agricultural Systems, vol. 76, pp. 1119--1135 (2003) 7. Passam, H. C., Sideridis, A. B., Yialouris, C. P., Maliappis, M. T.: Improvement of Vegetable Quality and Water and Fertilizer Utilization in Low-Tech Greenhouses through a Decision Support Management System. Journal of Vegetable Crop Production, vol. 7, pp. 69--82 (2001) 8. Simperl, E. P. B., Tempich C.: Ontology Engineering: a Reality Check. In: 5th International Conference on Ontologies, Databases, and Applications of Semantics ODBASE2006 (2006) 9. Soergel, D., Lauser, B., Liang, A., Fisseha, F., Keizer, J., Katz, S.: Reengineering Thesauri for New Applications: the AGROVOC Example. Journal of Digital Information, vol. 4 (2004)

498

M. T. Maliappis

10. Yialouris, C. P., Passam, H. C., Sideridis, A. B., Metin, C.: VEGES: A multilingual expert system for diagnosis of pests, diseases and nutritional disorders of six greenhouse vegetables. Computers and Electronics in Agriculture, vol. 19, pp. 55--67 (1997)